Merge f9654a3 into 4c3ccf1

sanity · May 21, 2015 · 70d52fa · 70d52fa
2 parents 4c3ccf1 + f9654a3
commit 70d52fa
Show file tree

Hide file tree

Showing 5 changed files with 60 additions and 12 deletions.
diff --git a/pom.xml b/pom.xml
@@ -32,7 +32,7 @@
         be accompanied by a bump in version number, regardless of how minor the change.
     -->
 
-    <version>0.7.13</version>
+    <version>0.7.14</version>
     <repositories>
         <repository>
             <id>sanity-maven-repo</id>

diff --git a/src/main/java/quickml/supervised/classifier/Classifiers.java b/src/main/java/quickml/supervised/classifier/Classifiers.java
@@ -24,8 +24,10 @@
 import quickml.supervised.predictiveModelOptimizer.PredictiveModelOptimizerBuilder;
 import quickml.supervised.predictiveModelOptimizer.fieldValueRecommenders.FixedOrderRecommender;
 
+import java.util.Arrays;
 import java.util.List;
 import java.util.Map;
+import java.util.Set;
 
 import static quickml.supervised.classifier.decisionTree.TreeBuilder.*;
 
@@ -64,8 +66,8 @@ public static <T extends ClassifierInstance> Pair<Map<String, Object>, Downsampl
         return new Pair<Map<String, Object>, DownsamplingClassifier>(bestParams, downsamplingClassifier);
     }
 
-    public static <T extends ClassifierInstance> Pair<Map<String, Object>, DownsamplingClassifier>  getOptimizedDownsampledRandomForest(List<T> trainingData, int rebuildsPerValidation, double fractionOfDataForValidation, ClassifierLossFunction lossFunction, DateTimeExtractor dateTimeExtractor, DownsamplingClassifierBuilder<T> modelBuilder) {
-        Map<String, FieldValueRecommender> config = createConfig();
+    public static <T extends ClassifierInstance> Pair<Map<String, Object>, DownsamplingClassifier>  getOptimizedDownsampledRandomForest(List<T> trainingData, int rebuildsPerValidation, double fractionOfDataForValidation, ClassifierLossFunction lossFunction, DateTimeExtractor dateTimeExtractor, DownsamplingClassifierBuilder<T> modelBuilder, Set<String> exemptAttributes) {
+        Map<String, FieldValueRecommender> config = createConfig(exemptAttributes);
         return getOptimizedDownsampledRandomForest(trainingData,  rebuildsPerValidation, fractionOfDataForValidation, lossFunction, dateTimeExtractor, modelBuilder, config);
     }
 
@@ -74,13 +76,12 @@ public static <T extends ClassifierInstance> Pair<Map<String, Object>, Downsampl
         return getOptimizedDownsampledRandomForest(trainingData,  rebuildsPerValidation, fractionOfDataForValidation, lossFunction, dateTimeExtractor, modelBuilder, config);
 
     }
-        public static Pair<Map<String, Object>, DownsamplingClassifier> getOptimizedDownsampledRandomForest(List<? extends ClassifierInstance> trainingData,  int rebuildsPerValidation, double fractionOfDataForValidation, ClassifierLossFunction lossFunction, DateTimeExtractor dateTimeExtractor) {
-        Map<String, FieldValueRecommender> config = createConfig();
+        public static Pair<Map<String, Object>, DownsamplingClassifier> getOptimizedDownsampledRandomForest(List<? extends ClassifierInstance> trainingData,  int rebuildsPerValidation, double fractionOfDataForValidation, ClassifierLossFunction lossFunction, DateTimeExtractor dateTimeExtractor, Set<String> exemptAttributes) {
+        Map<String, FieldValueRecommender> config = createConfig(exemptAttributes);
         return getOptimizedDownsampledRandomForest(trainingData,  rebuildsPerValidation, fractionOfDataForValidation, lossFunction, dateTimeExtractor, config);
     }
 
     private static int getTimeSliceHours(List<? extends ClassifierInstance> trainingData, int rebuildsPerValidation, DateTimeExtractor<ClassifierInstance> dateTimeExtractor) {
-
         Utils.sortTrainingInstancesByTime(trainingData, dateTimeExtractor);
         DateTime latestDateTime = dateTimeExtractor.extractDateTime(trainingData.get(trainingData.size()-1));
         int indexOfEarliestValidationInstance = (int) (0.8 * trainingData.size()) - 1;
@@ -91,14 +92,16 @@ private static int getTimeSliceHours(List<? extends ClassifierInstance> training
     }
 
 
-    private static  Map<String, FieldValueRecommender> createConfig() {
+    private static  Map<String, FieldValueRecommender> createConfig(Set<String> exemptAttributes) {
         Map<String, FieldValueRecommender> config = Maps.newHashMap();
         config.put(MAX_DEPTH, new FixedOrderRecommender(4, 8, 12));//Integer.MAX_VALUE, 2, 3, 5, 6, 9));
         config.put(MIN_OCCURRENCES_OF_ATTRIBUTE_VALUE, new FixedOrderRecommender(7, 10));
         config.put(MIN_LEAF_INSTANCES, new FixedOrderRecommender(0, 15));
         config.put(DownsamplingClassifierBuilder.MINORITY_INSTANCE_PROPORTION, new FixedOrderRecommender(.1, .25));
         config.put(DEGREE_OF_GAIN_RATIO_PENALTY, new FixedOrderRecommender(1.0, 0.75));
         config.put(ORDINAL_TEST_SPLITS, new FixedOrderRecommender(5, 7));
+        config.put(MIN_SPLIT_FRACTION, new FixedOrderRecommender(0.01, 0.25, .5 ));
+        config.put(EXEMPT_ATTRIBUTES, new FixedOrderRecommender(exemptAttributes));
         return config;
     }
 

diff --git a/src/main/java/quickml/supervised/classifier/decisionTree/TreeBuilder.java b/src/main/java/quickml/supervised/classifier/decisionTree/TreeBuilder.java
@@ -41,11 +41,15 @@ public final class TreeBuilder<T extends ClassifierInstance> implements Predicti
     public static final int RESERVOIR_SIZE = 100;
     public static final Serializable MISSING_VALUE = "%missingVALUE%83257";
     private static final int HARD_MINIMUM_INSTANCES_PER_CATEGORICAL_VALUE = 10;
+    public static final String MIN_SPLIT_FRACTION = "minSplitFraction";
+    public static final String EXEMPT_ATTRIBUTES = "exemptAttributes";
 
     private Scorer scorer;
     private int maxDepth = 5;
     private double minimumScore = 0.00000000000001;
     private int minDiscreteAttributeValueOccurances = 0;
+    private double minSplitFraction = .005;
+    private Set<String> exemptAttributes = Sets.newHashSet();
 
     private int minLeafInstances = 0;
 
@@ -74,6 +78,16 @@ public TreeBuilder attributeIgnoringStrategy(AttributeIgnoringStrategy attribute
         return this;
     }
 
+    public TreeBuilder exemptAttributes(Set<String> exemptAttributes) {
+        this.exemptAttributes = exemptAttributes;
+        return this;
+    }
+
+    public TreeBuilder minSplitFraction(double minSplitFraction) {
+        this.minSplitFraction = minSplitFraction;
+        return this;
+    }
+
     @Deprecated
     public TreeBuilder ignoreAttributeAtNodeProbability(double ignoreAttributeAtNodeProbability) {
         attributeIgnoringStrategy(new IgnoreAttributesWithConstantProbability(ignoreAttributeAtNodeProbability));
@@ -92,6 +106,8 @@ public TreeBuilder copy() {
         copy.ordinalTestSpilts = ordinalTestSpilts;
         copy.attributeIgnoringStrategy = attributeIgnoringStrategy.copy();
         copy.fractionOfDataToUseInHoldOutSet = fractionOfDataToUseInHoldOutSet;
+        copy.minSplitFraction = minSplitFraction;
+        copy.exemptAttributes = exemptAttributes;
         return copy;
     }
 
@@ -110,8 +126,12 @@ public void updateBuilderConfig(final Map<String, Object> cfg) {
             minCategoricalAttributeValueOccurances((Integer) cfg.get(MIN_OCCURRENCES_OF_ATTRIBUTE_VALUE));
         if (cfg.containsKey(MIN_LEAF_INSTANCES))
             minLeafInstances((Integer) cfg.get(MIN_LEAF_INSTANCES));
+        if (cfg.containsKey(MIN_SPLIT_FRACTION))
+            minSplitFraction((Double) cfg.get(MIN_SPLIT_FRACTION));
         if (cfg.containsKey(ORDINAL_TEST_SPLITS))
             ordinalTestSplits((Integer) cfg.get(ORDINAL_TEST_SPLITS));
+        if (cfg.containsKey(EXEMPT_ATTRIBUTES))
+            exemptAttributes((Set<String>) cfg.get(EXEMPT_ATTRIBUTES));
         if (cfg.containsKey(DEGREE_OF_GAIN_RATIO_PENALTY))
             degreeOfGainRatioPenalty((Double) cfg.get(DEGREE_OF_GAIN_RATIO_PENALTY));
         if (cfg.containsKey(ATTRIBUTE_IGNORING_STRATEGY))
@@ -441,6 +461,12 @@ private Pair<? extends Branch, Double> createTwoClassCategoricalNode(Node parent
             inCounts = inCounts.add(testValCounts);
             outCounts = outCounts.subtract(testValCounts);
 
+            double numInstances = inCounts.getTotal() + outCounts.getTotal();
+            if (!exemptAttributes.contains(attribute) && (inCounts.getTotal()/ numInstances <minSplitFraction ||
+                    outCounts.getTotal()/ numInstances < minSplitFraction)) {
+                continue;
+            }
+
             if (inCounts.getTotal() < minLeafInstances || outCounts.getTotal() < minLeafInstances) {
                 continue;
             }
@@ -666,6 +692,12 @@ private Pair<? extends Branch, Double> createNumericBranch(Node parent, final St
             ClassificationCounter inClassificationCounts = ClassificationCounter.countAll(inSet);
             ClassificationCounter outClassificationCounts = ClassificationCounter.countAll(outSet);
 
+            double numInstances = inClassificationCounts.getTotal() + outClassificationCounts.getTotal();
+            if (!exemptAttributes.contains(attribute) && (inClassificationCounts.getTotal()/ numInstances <minSplitFraction ||
+                    outClassificationCounts.getTotal()/ numInstances < minSplitFraction)) {
+                continue;
+            }
+
             if (binaryClassifications) {
                 if (attributeValueOrIntervalOfValuesHasInsufficientStatistics(inClassificationCounts)
                         || inClassificationCounts.getTotal() < minLeafInstances
@@ -683,7 +715,7 @@ private Pair<? extends Branch, Double> createNumericBranch(Node parent, final St
             if (thisScore > bestScore) {
                 bestScore = thisScore;
                 bestThreshold = threshold;
-                probabilityOfBeingInInset = inClassificationCounts.getTotal() / (inClassificationCounts.getTotal() + outClassificationCounts.getTotal());
+                probabilityOfBeingInInset = inClassificationCounts.getTotal() / numInstances;
             }
         }
         if (bestScore == 0) {

diff --git a/...rvised/classifier/StaticBuildersTest.java → ...upervised/classifier/ClassifiersTest.java b/...rvised/classifier/StaticBuildersTest.java → ...upervised/classifier/ClassifiersTest.java
@@ -1,5 +1,6 @@
 package quickml.supervised.classifier;
 
+import com.beust.jcommander.internal.Sets;
 import org.javatuples.Pair;
 import org.junit.Test;
 import org.slf4j.Logger;
@@ -9,23 +10,29 @@
 import quickml.supervised.classifier.downsampling.DownsamplingClassifier;
 import quickml.supervised.crossValidation.lossfunctions.WeightedAUCCrossValLossFunction;
 
+import java.util.Arrays;
 import java.util.List;
 import java.util.Map;
+import java.util.Set;
 
 import static quickml.supervised.InstanceLoader.getAdvertisingInstances;
 
 
-public class StaticBuildersTest {
-    private static final Logger logger = LoggerFactory.getLogger(StaticBuildersTest.class);
+public class ClassifiersTest {
+    private static final Logger logger = LoggerFactory.getLogger(ClassifiersTest.class);
+
 
 @Test
     public void getOptimizedDownsampledRandomForestIntegrationTest() throws Exception {
         double fractionOfDataForValidation = .2;
         int rebuildsPerValidation = 1;
-        List<ClassifierInstance> trainingData = getAdvertisingInstances().subList(0, 3000);
+        Set<String> exemptAttributes = Sets.newHashSet();
+        exemptAttributes.addAll(Arrays.asList("seenClick", "seenCampaignClick", "seenPixel", "seenCampaignPixel", "seenCreativeClick", "seenCampaignClick"));
+
+    List<ClassifierInstance> trainingData = getAdvertisingInstances().subList(0, 3000);
         OnespotDateTimeExtractor dateTimeExtractor = new OnespotDateTimeExtractor();
         Pair<Map<String, Object>, DownsamplingClassifier> downsamplingClassifierPair =
-                Classifiers.<ClassifierInstance>getOptimizedDownsampledRandomForest(trainingData, rebuildsPerValidation, fractionOfDataForValidation, new WeightedAUCCrossValLossFunction(1.0), dateTimeExtractor);
+                Classifiers.<ClassifierInstance>getOptimizedDownsampledRandomForest(trainingData, rebuildsPerValidation, fractionOfDataForValidation, new WeightedAUCCrossValLossFunction(1.0), dateTimeExtractor, exemptAttributes);
         logger.info("logged weighted auc loss should be between 0.25 and 0.28");
     }
 }
diff --git a/.../quickml/supervised/predictiveModelOptimizer/PredictiveModelOptimizerIntegrationTest.java b/.../quickml/supervised/predictiveModelOptimizer/PredictiveModelOptimizerIntegrationTest.java
@@ -54,6 +54,9 @@ public void testOptimizer() throws Exception {
     private Map<String, FieldValueRecommender> createConfig() {
         Map<String, FieldValueRecommender> config = Maps.newHashMap();
         Set<String> attributesToIgnore = Sets.newHashSet();
+        Set<String> exemptAttributes = Sets.newHashSet();
+        exemptAttributes.addAll(Arrays.asList("seenClick", "seenCampaignClick", "seenPixel", "seenCampaignPixel", "seenCreativeClick", "seenCampaignClick"));
+
         attributesToIgnore.addAll(Arrays.asList("browser", "eap", "destinationId", "seenPixel", "internalCreativeId"));
         double probabilityOfDiscardingFromAttributesToIgnore = 0.3;
         CompositeAttributeIgnoringStrategy compositeAttributeIgnoringStrategy = new CompositeAttributeIgnoringStrategy(Arrays.asList(
@@ -67,6 +70,9 @@ private Map<String, FieldValueRecommender> createConfig() {
         config.put(MIN_LEAF_INSTANCES, new FixedOrderRecommender(0, 20, 40));
         config.put(SCORER, new FixedOrderRecommender(new InformationGainScorer(), new GiniImpurityScorer()));
         config.put(DEGREE_OF_GAIN_RATIO_PENALTY, new FixedOrderRecommender(1.0, 0.75, .5 ));
+        config.put(MIN_SPLIT_FRACTION, new FixedOrderRecommender(0.01, 0.25, .5 ));
+        config.put(EXEMPT_ATTRIBUTES, new FixedOrderRecommender(exemptAttributes));
+
         return config;
     }